pokered
pokered copied to clipboard
Automatically detect invalid object_event or bg_event text IDs
See https://github.com/pret/pokered/pull/367#issuecomment-1181936172
Since we already have def_warps_to
following the bg and object event definitions, it can be responsible for the checking:
MACRO bg_event
db \2, \1, \3
+ DEF _BG_EVENT_{d:{_NUM_BG_EVENTS}}_TEXT_ID = \3
DEF {_NUM_BG_EVENTS} += 1
ENDM
MACRO object_event
...
+ ; items and trainers don't use a typical text id
+ IF _NARG > 6
+ DEF _OBJECT_EVENT_{d:{_NUM_OBJECT_EVENTS}}_TEXT_ID = 0
+ ELSE
+ DEF _OBJECT_EVENT_{d:{_NUM_OBJECT_EVENTS}}_TEXT_ID = \6
+ ENDC
DEF {_NUM_OBJECT_EVENTS} += 1
ENDM
MACRO def_warps_to
+ ; text ID values are significant (see DisplayTextID in home/text_script.asm)
+ FOR n, {_NUM_BG_EVENTS}
+ ASSERT _BG_EVENT_{d:n}_TEXT_ID > {_NUM_OBJECT_EVENTS}, \
+ "A bg_event has text ID {d:_BG_EVENT_{d:n}_TEXT_ID} expected for an object_event ({d:{_NUM_OBJECT_EVENTS}} or below)"
+ ENDR
+ FOR n, {_NUM_OBJECT_EVENTS}
+ ASSERT _OBJECT_EVENT_{d:n}_TEXT_ID <= {_NUM_OBJECT_EVENTS}, \
+ "An object_event has text ID {d:_OBJECT_EVENT_{d:n}_TEXT_ID} expected for a bg_event (above {d:{_NUM_OBJECT_EVENTS}})"
+ ENDR
- FOR n, _NUM_WARP_EVENTS
+ FOR n, {_NUM_WARP_EVENTS}
warp_to _WARP_{d:n}_X, _WARP_{d:n}_Y, \1_WIDTH
ENDR
ENDM
This would go well with how warp_event
defines per-event _WARP_{d:{_NUM_WARP_EVENTS}}_X
and _WARP_{d:{_NUM_WARP_EVENTS}}_Y
constants which get used by def_warps_to
.
This should be documented as a design flaw (in the wiki since we don't yet have a definitive document worth putting in the repo), which could be fixed by editing DisplayTextID
in home/text_script.asm to check for "sprite-ness" another way. (Not sure how though.)
This should be documented as a design flaw (in the wiki since we don't yet have a definitive document worth putting in the repo), which could be fixed by editing
DisplayTextID
in home/text_script.asm to check for "sprite-ness" another way. (Not sure how though.)
Whatever the case, a design flaws wiki page now exists for those willing to put it in: https://github.com/pret/pokered/wiki/Design-Flaws
@Rangi42 Your suggested changes causes a conflict with the constants being created in my PR:
error: maps.asm(255) -> data/maps/objects/LavenderTown.asm(16) -> macros/scripts/maps.asm::bg_event(69):
Expected constant expression: 'TEXT_LAVENDERTOWN_POKECENTER_SIGN' is not constant at assembly time
Could this be because the macros are being called before the constants are defined?
Yeah, the problem is DEF constant = value
needs value
to be known at assembly time, but if you have bg_event 4, 5, TEXT_LAVENDERTOWN_POKECENTER_SIGN
, then the macro tries to do DEF _BG_EVENT_3_TEXT_ID = TEXT_LAVENDERTOWN_POKECENTER_SIGN
and doesn't have a value for TEXT_LAVENDERTOWN_POKECENTER_SIGN
yet (since that gets defined by the const_dw
table in its script file).
Using EQUS
might work instead:
MACRO bg_event
db \2, \1, \3
+ REDEF _BG_EVENT_{d:{_NUM_BG_EVENTS}}_TEXT_ID EQUS "\3"
DEF {_NUM_BG_EVENTS} += 1
ENDM
MACRO object_event
...
+ ; items and trainers don't use a typical text id
+ IF _NARG > 6
+ REDEF _OBJECT_EVENT_{d:{_NUM_OBJECT_EVENTS}}_TEXT_ID EQUS "0"
+ ELSE
+ REDEF _OBJECT_EVENT_{d:{_NUM_OBJECT_EVENTS}}_TEXT_ID EQUS "\6"
+ ENDC
DEF {_NUM_OBJECT_EVENTS} += 1
ENDM
MACRO def_warps_to
+ ; text ID values are significant (see DisplayTextID in home/text_script.asm)
+ FOR n, {_NUM_BG_EVENTS}
+ ASSERT {_BG_EVENT_{d:n}_TEXT_ID} > {_NUM_OBJECT_EVENTS}, \
+ "A bg_event has a text ID {_BG_EVENT_{d:n}_TEXT_ID} expected for an object_event ({d:{_NUM_OBJECT_EVENTS}} or below)"
+ ENDR
+ FOR n, {_NUM_OBJECT_EVENTS}
+ ASSERT {_OBJECT_EVENT_{d:n}_TEXT_ID} <= {_NUM_OBJECT_EVENTS}, \
+ "An object_event has a text ID {_OBJECT_EVENT_{d:n}_TEXT_ID} expected for a bg_event (above {d:{_NUM_OBJECT_EVENTS}})"
+ ENDR
- FOR n, _NUM_WARP_EVENTS
+ FOR n, {_NUM_WARP_EVENTS}
warp_to _WARP_{d:n}_X, _WARP_{d:n}_Y, \1_WIDTH
ENDR
ENDM
That worked and is functional with the changes in my new PR (Tested by purposely changing around some TEXT constants). I'm going to incorporate this into my PR and will push an update a bit later.